ALTK-Evolve learned memory tops τ²-bench customer service leaderboard
#1 on τ²-bench
Tops τ²-bench customer service leaderboard without a frontier-class model
Researchers reported that ALTK-Evolve, a learned-memory system for AI agents, reached the top of the τ²-bench customer service leaderboard. The result came without depending on a frontier-class model, pointing to memory design as a lever for agent performance.
Read the full story at Hugging Face
The tools behind this win
- ALTK-EvolveLearned-memory approach for AI agents that topped the benchmark